Measuring Organizational Strategies: Some Theoretical and Methodological Problems

Author(s): Charles C. Snow and Donald C. Hambrick

Source: The Academy of Management Review , Oct., 1980, Vol. 5, No. 4 (Oct., 1980), pp.
527-538
Published by: Academy of Management

Stable URL: https://www.jstor.org/stable/257458

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms

Academy of Management is collaborating with JSTOR to digitize, preserve and extend access to
The Academy of Management Review

This content downloaded from

210.57.217.22 on Fri, 11 Nov 2022 01:37:31 UTC
All use subject to https://about.jstor.org/terms
Academy of Management Review 1980, Vol. 5, No. 4, 527-538

Measuring Organizational Strategies:

Some Theoretical and
Methodological Problems
CHARLES C. SNOW
Pennsylvania State University

DONALD C. HAMBRICK
Columbia University
In this article we address the major theoretical and methodological prob-
lems encountered in attempts to arrive at valid and reliable measures of
organizational strategy. Our discussion is based on a series of empirical
studies of the strategic behaviors of nearly 200 organizations in ten in-
dustries. In these studies, four different approaches for measuring
strategy have been employed. We describe each approach and discuss
its advantages and disadvantages.

In order to survive and prosper, every organiza- The paper is divided into four major sections.
tion must develop and maintain an acceptable The first is a brief historical overview, describing
alignment with its environment. Increasingly,the evolution of the strategy concept. The second
organization and management theorists are view-section is a discussion of recurring theoretical
ing strategy as the mechanism that guides en-problems in the area of strategy, supported with
vironmental alignment and provides integration for examples from our own and others' research. In
internal operations. the third section, we discuss four approaches to
Over the past several years, we and several col-measuring strategy, pointing out their respective
leagues have investigated the role of strategy in strengths and weaknesses in relation to the theo-
organizational behavior and analysis. During the retical issues raised in the preceding section. The
course of six major research projects using a varie-final section presents our conclusions.
ty of measurement techniques, we have faced
many difficult theoretical and methodological The Concept of Strategy:
problems in attempting to arrive at valid and Origin and Definition
reliable measures of organizational strategy. Our The word strategy is derived from the Greek
purpose in this article is to describe and discuss strategos - literally, "the art of the general" [Hart,
the major problems encountered to date so that 1967]. The concept of strategy apparently was
future investigators will be apprised of some im-introduced into the organizational literature and
portant and recurring issues involved in the advanced most notably during the fifties by faculty
measurement of strategy. (For a more detailed members at the Harvard Business School. The
discussion of the research on which this article is
Harvard view of strategy was (and is) normative, in
based, see Miles and Snow [1978]; Snow and that strategy was treated as a situational art, an im-
Hrebiniak [1980]; and Hambrick [1979].) aginative act of integrating numerous complex
decisions [Andrews, Learned, Christensen, &
'We wish to thank David Jemison for his helpful comments on
an earlier version of this article.
Guth, 1965; Andrews, 1971].
? 1980 by the Academy of Management 0363-7425 In contrast to this normative approach,
527

This content downloaded from

210.57.217.22 on Fri, 11 Nov 2022 01:37:31 UTC
All use subject to https://about.jstor.org/terms
Chandler's research was the first to employ nearly 200 organizations in ten industries. This
strategy as a descriptive concept. From his study research has focused on what Hofer and Schendel
of four companies that pioneered in diversification, [1978] have called business-level strategies ("How
Chandler concluded that strategy was the key should we compete in this business?"), as op-
mechanism used for charting a new direction, and posed to corporate-level strategies ("What
its impact on organizational structure and perfor- businesses should we be in?"). Moreover, this
mance was substantial. In Chandler's view, research has dealt with the content rather than the
strategy refers to "the determination of the basic process of strategy [Hofer, 1975]. That is, the focus
long-term goals and objectives of the enterprise has been primarily on strategy and its correlates,
and the adoption of courses of action and the al- not the processes used to formulate and imple-
location of resources necessary for carrying out ment strategies. The individual studies have
these goals" [1962, p.13]. Clearly, this definition of employed a variety of methods, including intensive
strategy includes elements of both ends (objec- interviewing, mailed questionnaire surveys, and
tives and goals) and means (courses of action and combinations of interviews and questionnaires.
allocation of resources). Today, strategy theorists Each of the studies has generated its own unique
do not agree about whether the concept should in- set of theoretical and methodological problems,
clude goals and objectives as well as the means but some of these problems have tended to recur.
used to achieve them [Hofer & Schendel, 1978]. It is not possible to provide definitive answers to
Researchers have, however, reached a general these problems based on our research. However,
consensus on distinguishing between strategy for- in the discussion below we will attempt to link our
mulation and strategy implementation. The advan- experiences with those of other strategy scholars
tage of making this distinction is that the cognitive in discussing the following issues: (1) What con-
aspects of strategy (formulation) can be viewed as stitutes a change in strategy, as opposed to a
an important phase apart from the action compo- strategic adjustment? (2) Is there a difference be-
nent (implementation). This formulation/ tween intended and realized strategies? How do
implementation dichotomy is useful concep- these different forms of strategy arise? How can
tually, but it implies that strategy is developed con- the investigator recognize which type of strategy is
sciously and purposefully. Although this may be being observed? (3) What are some of the
true in many instances, Mintzberg [1978] has problems that can occur when industries and or-
described how organizational strategies can ganizations are being considered for inclusion in a
emerge unintentionally. Therefore, in order to in- research sample?
clude both premeditated and emergent strategies
in theoretical conceptualizations, Mintzberg as Strategic Change and Adjustment
well as Miles and Snow [1978] have suggested that Strategies have been described as changing
researchers should view strategy as a pattern in over time, sometimes gradually and incrementally
the organization's important decisions and ac- [Quinn, 1978], sometimes quickly and radically
tions. Typically, these decisions will be directed at [Mintzberg, 1978]. At first glance, an organization's
(1) maintaining the organization's alignment with response to environmental change might be
its environment and (2) managing its major internal viewed as a strategic change. However, if an
interdependencies. Defining strategy in this man- organization responds to a particular environmen-
ner allows researchers to move beyond the ab- tal change in a familiar way, then this response is
stract and normative aspects of strategy toward probably a continuation of, not a change in,
those decisions which actually involve organiza- strategy. For example, during the mid-1970s one of
tional goals and the allocation of resources the college textbook publishing firms in our
necessary to achieve goals. studies was actively pursuing opportunities in the
relatively new and expanding continuing-
education market. This company's response to
Major Theoretical Problems
previous market opportunities had been similarly
At this writing, we and our associates have con- eager and aggressive over a period of ten to fifteen
ducted six studies of the strategic behaviors of years. Thus the firm qualified as a Prospector (in
528

This content downloaded from

210.57.217.22 on Fri, 11 Nov 2022 01:37:31 UTC
All use subject to https://about.jstor.org/terms
the Miles-Snow [1978] terminology), and its in- it is still a Defender, but with an enlarged domain.
itiative in the continuing-education market The distinction between change and adjustment
amounted to a continuation of the Prospector may be subtle, but it is important to theory
strategy. Of course, researchers who view strategy building. If most or all domain and structural
in terms of niche selection might conclude for changes are regarded as strategic changes, then
their purposes that the firm had changed its theories can only be fragmented and ephemeral.
strategy. But if these same modifications are viewed as ad-
Most organizations prefer to apply familiar solu- justments, researchers can then look for con-
tions to new problems [Cyert & March, 1963; sistencies in how an organization interacts with its
Cohen, March, & Olsen, 1972]. Our research sup- environment. Mintzberg [1978], for example, has of-
ports this view, suggesting that organizations are fered the concept of metastrategy - an underly-
generally reluctant to change their strategies. The ing, enduring orientation that an organization has
investment in time, people, money, and other re- in relation to its environment. Metastrategies may
sources required to develop the distinctive com- be adjusted as they evolve (through lengthy life
petencies, technologies, structures, and manage- cycles of conception, elaboration, decay, and
ment processes needed to pursue a particular death), but they are changed infrequently.
strategy is large. In addition, the managerial stress Of the few strategic changes we have observed,
associated with planning and executing strategic the most rapid took place in a textbook firm. Over a
change is often a deterrent to major or frequent six-month period, this company changed from a
changes in strategy. Perhaps the greatest obstacle Reactor to an Analyzer (in the Miles-Snow ter-
to strategic change stems from the fact that over minology), including the technological and struc-
time a given strategy attracts and fosters a set of tural modifications that management made to fit
managerial values and philosophies that are wed- the new strategy. We believe that strategic change
ded to the strategy [Beer & Davis, 1976; Guth & cannot be dissociated from the technological,
Taguiri, 1965; Richards, 1973]. Thus, if at all possi- structural, and process changes required to effect-
ble, organizations, when faced with external ively pursue a new strategy. And, in fact, a useful
change or pressure, tend to adjust rather than view might be that strategic change occurs only
change their strategies. when the organization (1) modifies in a major way
its alignment with the environment and (2) substan-
The experience of one of the colleges in our
tially alters technology, structure, and process to
studies exemplifies the problems associated with
fit the new alignment. Until an organization alters
trying to determine whether or not an organization
its internal features to support a strategic
has made a strategic change. Prior to 1972, the col-
modification, it is speculative to conclude that the
lege could be characterized as a Defender: its own
modification is anything more than a strategic ad-
administrators, a panel of higher education ex-
justment. For example, of the sixteen textbook
perts, and a set of objective indicators all indicated
companies studied by Miles and Snow [1978], only
that the college offered an unchanging curriculum
three had changed their strategies (and accompa-
to a relatively stable student market. Faced with
nying structures) over the previous six years, but all
rapidly declining enrollments and severe financial
sixteen had made adjustments to their strategies.
problems in 1971-72, the college responded over
In sum, it is important for researchers interested
the next four years by adding various new pro-
in strategy to be aware of the fact that the general
grams to its curriculum (an MBA program, health
tendency of managers is to preserve rather than
care programs, etc.). During the period 1977-79, the
change their organizations' strategies. Because
college, in one administrator's words, was in the
strategic change and adjustment is frequently a
process of "settling down" - developing no new
lengthy process, investigators must attempt to
programs and incorporating the recent program
determine the point at which they have entered the
additions into the organization's structure.
strategic life cycle. The question becomes, "Am I
Did this organization undergo a strategic measuring past strategy, a new strategy, a period
change? We believe not. Rather, over this seven- of strategic adjustment, or some conception of
year period, the organization adjusted its strategy; future strategy?"
529

This content downloaded from

210.57.217.22 on Fri, 11 Nov 2022 01:37:31 UTC
All use subject to https://about.jstor.org/terms
 Intended and Realized Strategies perhaps because no strategy was intended at
the outset or perhaps because those that were
Granted that it is difficult to ascertain whether or intended got displaced along the way; these
not an organization's strategy has changed, it is may be called emergent strategies [adapted
frequently even more difficult to determine from Mintzberg, 1978, p. 945].
whether the organization has a strategy [Mintzberg, Our research permits us to elaborate on and ex-
1978; Quinn, 1977]. Is a strategy that management tend Mintzberg's discussion of intended,
has formulated but not fully implemented really a emergent, and realized strategies. We have en-
strategy? Or, conversely, is an apparent "pattern in countered a variety of situations where it was dif-
a stream of important decisions" that was not ficult to determine an organization's intended
planned by management actually a strategy? For strategy, but three different explanations seem to
some researchers, this distinction between inten- account for the majority of these situations.
tion and realization may not be critical. For others, Probably the most significant reason that resear-
it will be of central importance. Strategies are chers have trouble discerning organizations'
formed by people, and unless these strategies are strategies is that managers seldom conceive of
communicated fully, the researcher may be tap- strategy in the same terms as researchers. For ex-
ping into some unknowable combination of stra- ample, the Miles-Snow strategic typology is based
tegic history, managerial desires, and plans largely on organizations' orientations toward
[Mintzberg, 1978]. An interesting and useful product-market development. Managers, however,
framework for conceptualizing intended and typically do not think of their organizations as be-
realized strategies has been offered by Mintzberg ing Defenders, Prospectors, Analyzers, or Reac-
and is diagrammed below: tors. Instead, they may think of their organizations'
strategies as resulting from concerns about being
biggest, best, first, lowest priced, highest quality,
and so forth. Therefore, a large part of the problem
Intended Deliberate Realized researchers and managers have in communicating
Strategy Strategy - Strategy with each other is language. Labels like Prospector
/ and Defender aid in conceptualization, but they do
not necessarily capture the nature of strategy as
Unrealized Emergent
Strategy Strategy seen by those who formulate and implement it.
Second, research by Quinn [1977, 1978] in-
(After Mintzberg, H. Patters in strategy formation. dicates that top managers often avoid announcing
Management Science, 1978, 24, p. 945.
explicit goals and, in turn, articulating strategy.
Rather, they keep their options open as long as
Figure 1
possible so that they do not close off issues that
Intended and Realized Strategies
are important to subordinates, so that they do not
provide a target for opposition, so that corporate
security is not breached, and so on. Researchers
Using illustrations from historical analyses of who enter the organization looking for goals and
Volkswagenwerk and of the U.S. government in strategies may interpret this vagueness as an
Vietnam, Mintzberg discussed various combina- absence of strategy, whereas Quinn [1977] believes
tions of intended and realized strategies: that it is politically astute and effective manage-
1. Intended strategies that are realized; these may
ment. In any event, this form of goal setting by top
be called deliberate strategies. managers makes it difficult to learn an organiza-
tion's intended strategy.
2. Intended strategies that are not realized, Finally, we have observed a few organizations
perhaps because of unrealistic expectations,
that did not appear to have any intended strategy.
misjudgments about the environment, or
changes in either during implementation; Usually, this turns out to be a temporary condition.
these may be called unrealized strategies. However, in some cases, the absence of a clear, in-
tegrated strategy may be prolonged. Some
3. Realized strategies that were never intended, organizations can weather environmental threats
530

This content downloaded from

210.57.217.22 on Fri, 11 Nov 2022 01:37:31 UTC
All use subject to https://about.jstor.org/terms
for considerable periods without a solid strategy present service level while searching for cost effi-
because of slack resources. Some exist in relative- ciencies.
ly benign environments that do not require well- Another hospital provides an example of how an
conceived strategies - e.g., airline companies emergent strategy can stem from internal political
before deregulation [see Snow & Hrebiniak, 1980]. forces. This hospital had a formal plan that called
The realized strategies of "The Headless Giant" for controlled expansion of selected departments
and "The Stagnant Bureaucracy" identified by within the hospital over a five-year period. The
Miller and Friesen [1978] certainly could not have medical staff in the departments slated for no ex-
been the intended strategies of their top manage- pansion coalesced and, through a series of
ment groups. In all of the situations just described, political maneuvers, secured capital expenditures
the researcher who is attempting to determine the for their units equal to the planned two-year capital
organization's intended strategy is faced with a dif- expenditure budget for the entire hospital. In this
ficult task. case, the hospital intended to pursue one strategy
We have also encountered emergent strategies. but another emerged through purely political pro-
In the main, these strategies appeared to be driven cesses.

by unforeseen threats, opportunities, or internal In sum, researchers should have a cle

political forces. One textbook firm found itself with whether their purposes are best served
a product-market domain that was the exact op- ing intended or realized strategies. If the
posite of the areas where growth was occurring. terest is in realized strategies, researc
Because the growth areas were largely outside its look further to determine whether these
familiar arena, the organization was reluctant to are the result of management inten
move in any direction aggressively. Finally, faced emergent processes.
with rapidly increasing inventories and declining
sales, top management replaced the college divi-
The Relativity of Strategy
sion director with the national sales manager. This
individual quickly proceeded to move the division The task of measuring strategy is to
toward the growth areas by consolidating some solute values to what is, in fact, a relative
units, altering the structure of others, and phenomenon. An organization's business-level
reconstituting the mix of key personnel. Today, as strategy is largely predicated on industry condi-
a result of the changes brought on by its misalign- tions and competitors' actions [e.g., Caves and
ment with the environment, this company exhibits Porter, 1977; Hofer, 1975]. We have encountered,
a well-conceived strategy that has produced primarily through the sampling process, several
desired profit levels. unanticipated problems in our research that have
One of the hospitals in our research had an stemmed from the relativity of the strategies
emergent strategy that evolved as a result of en- observed. The most important problem concerns
vironmental opportunities. This hospital was the degree of homogeneity/heterogeneity in the in-
located in a rapidly growing suburban area. The dustry. For example, Hambrick [1979] attempted to
area's population was composed largely of upper- obtain a sample of "pure" Defenders and Prospec-
middle-class professionals and their families, who tors from among a group of private liberal arts col-
demanded convenient, high-quality medical care. leges. In order to control for other key variables, he
Over a ten-year period, the hospital administration selected from medium-sized (1000-5000 students)
continually added new programs and services schools in the middle of the student quality range
without much concern about cost or integration of (SAT scores of 1000-1200). During later interviews,
services. During the researcher's visits to this it became apparent that this sampling approach
organization, no intended strategy was apparent may have restricted the range of strategic behavior
(unless "adding every service requested" could be observed. If the sample had included very high
considered an intended course of action). Today, quality private colleges, more extreme Defender
however, the hospital and the community essen- orientations might have been observed [Clark,
tially have stopped growing, and management has 1972]. Or, if community colleges had been includ-
a clear strategy for maintaining the hospital's ed, more extreme Prospector organizations might
531

This content downloaded from

210.57.217.22 on Fri, 11 Nov 2022 01:37:31 UTC
All use subject to https://about.jstor.org/terms
have been located [Blocker, Bender, & Martorana, focus only on strategic situations that were new,
1975]. dramatic, or otherwise newsworthy. More mun-
Sample heterogeneity can also be as much of a dane strategic behavior within the same industries
problem as homogeneity. In a study of electronics and organizations may have gone unnoticed.
and food-processing companies, Coleman ran- A different sort of problem occurred when Ham-
domly selected firms in each of these industries brick [1979] attempted to use secondary sources to
[Miles & Snow, 1978, Chap. 12]. Part of the process corroborate perceptual measures of present
for measuring strategy involved asking top strategy. In every instance, the most current data
managers to characterize the strategies of their sources were one to two years out of date, and in
competitors. In each industry, however, some the meantime some of the organizations had made
managers were unable to comment on their com- strategic adjustments or changes.
petitors' strategic behavior because these firms In sum, the nature of the sample of organiza-
operated in quite different segments of the in- tions selected for study can create theoretical or
dustry. Therefore, the lack of information about a measurement problems owing to the relativity of
given company precluded an accurate measure- strategy. Attention at the outset of a study to the
ment of strategy in several cases. As a general rule issues of industry homogeneity/heterogeneity,
for sampling within an industry, it should be comparability of measures across industry, and
assumed that the industry will contain a number of potential problems with secondary data sources
segments [Caves & Porter, 1977; Hatten & can reduce many troubling problems later in the
research.
Schendel, 1977]. All of the organizations in the in-
dustry will not operate in all of these segments.
Measurement Approaches
Any theory of organizational strategy would ap-
pear to be enhanced if it generalized across in- Researchers have available an array of options
dustries. However, moving from one industry to for identifying and measuring strategies. Our
another may cause problems in terms of com- research on business-level strategy has employed
parability among measures of strategy and its cor- four different approaches. Used in combination,
relates. For example, Hambrick [1979] had no these techniques usually converge to portray a par-
assurance that the Prospector colleges in his sam- ticular strategy for the organization in question.
ple were located in the same range as the Prospec- Used independently, each method has its own ad-
tor hospitals on the hypothetical absolute scale of vantages and disadvantages. The four measure-
strategic orientation. If a theory is tested under ment approaches are: (1) investigator inference, (2)
such circumstances and yields mixed results self-typing, (3) external assessment, and (4) objec-
across industries, it is difficult to conclude tive indicators. Their strengths and weaknesses,
whether the theory is only selectively valid or relative to the theoretical issues just discussed,
whether the strategy measurements differed are presented in Table 1.
across industries. No simple solution to this
problem exists. At this stage of theory develop- Investigator Inference
ment, it would be wise to conduct careful pilot In this approach, the researcher, using all the in-
studies of organizations in the different industries formation available, assesses the organization's
before a final sampling plan is prepared. strategy. A favorite method of case writers, this
A final problem has to do with secondary data measurement approach has few advantages and
sources. Such sources, if used exclusively, can numerous shortcomings. Clearly, the primary ad-
restrict observations of different organizational vantage is that the researcher may have more infor-
strategies. For example, the business cases mation about the organization's operations than
analyzed by Miller and Friesen [1978] included only anyone except the managers themselves. Combin-
"highly successful" and "unsuccessful" firms ing this information with a theoretical framework,
whose behavior had engaged the attention and and the ability to see the organization more objec-
predilections of case and magazine writers. The tively than its managers, may allow the investigator
potential clearly existed for these observers to to make a relatively accurate identification of
532

This content downloaded from

210.57.217.22 on Fri, 11 Nov 2022 01:37:31 UTC
All use subject to https://about.jstor.org/terms
strategy. tion's typical pattern of behavior over time rather
The primary disadvantage is that samples large than for any specific period.
enough to provide sufficient data for testing Although this approach is useful for measuring
hypotheses seldom can be generated using this the strategies of a large number of organizations in
approach. If the investigator is required to obtain a given study, it also has several shortcomings.
all of the information needed to understand what First, many managers believe that their organiza-
strategic decisions have been and are being made, tion is unique, and therefore they resist attempts to
then the number of organizations that can be classify it. Meyer found that hospital admin-
studied is likely to be very limited. Other problems istrators were both willing and able to classify their
with this approach involve the investigator's own competitor hospitals according to strategic orien-
perceptual biases, potentially limited access to tation [Miles & Snow, 1978, Chap. 13]. However,
key decisions (especially planned strategic these same administrators often were reluctant to
changes), a tendency to interpret the or- classify their own hospitals because of their
ganization's behavior according to an implicit perceived uniqueness. Apparently, many man-
theory, and so on. On balance, we have found the agers see so much differentiation and complexity
method of investigator inference to be a relatively in their organizations that they are unable to use
weak measurement approach. It is often less ac- strategic typologies easily. Second, Snow and
curate than the approaches described below, and Hrebiniak [1980] found some variance among
it can be applied only to relatively small samples of managers' perceptions of strategy within organ-
organizations. izations. Although most of the 88 firms in their
study appeared to have a good deal of managerial
Self-Typing consensus on strategy, there was enough variance
in perceptions to indicate that this measurement
An alternative to investigator inference is to approach is not infallible. Bourgeois [1978] also
allow the organization's managers (specifically, its found substantial dissent regarding strategy
top managers) to characterize the organization's within some of the 12 organizations he studied.
strategy. This approach was used by Coleman Third, Hambrick [1979] concluded that executives
[Miles & Snow, 1978, Chap. 12] and by Snow and generally report their organizations' intended
Hrebiniak [1980]. Using written descriptions of the strategies, as opposed to emergent or realized
four strategies in the Miles-Snow typology, top strategies. In those instances when no intended
managers in both studies were asked to classify strategy exists, an executive may create one for
their own organization as a Defender, Prospector, the benefit of the researcher. This is a common
Analyzer, or Reactor. The advantages of this problem in the social sciences [Nisbett & Wilson,
method are: (1) top managers' perceptions and 1977]. Finally, asking only the organization's
opinions largely determine the organization's managers to assess strategy does not permit any
strategy (at least its intended strategy), and (2) external or objective confirmation of their answers.
relatively large data bases can be generated for Therefore, even though there may be general
hypothesis testing. agreement within top-management teams about
Coleman, in a study of 27 food-processing com- strategy, self-typing shares with the investigator-
panies and 22 electronics firms, supplied written inference approach a lack of external confirmation.
descriptions of the strategy types to managers dur-
ing an interview, whereas Snow and Hrebiniak in- Exteral Assessment
cluded similar descriptions as part of a lengthy
mailed questionnaire (in neither case were the In three of our studies, we have attempted t
terms Defender, Prospector, etc., used). In both the confirm self-typing measures of strategy by obta
oral and written modes, the instructions to the ing the ratings of individuals external to the f
respondent were similar: (1) use industry com- organization (e.g., competitors, consultants, in
petitors as a frame of reference, (2) consider the dustry analysts, and expert panels). As mention
organization as a whole (or the division in a divi- earlier, Coleman found that managers in both
sionalized company), and (3) think of the organiza- food-processing and electronics samples
533

This content downloaded from

210.57.217.22 on Fri, 11 Nov 2022 01:37:31 UTC
All use subject to https://about.jstor.org/terms
 Table 1
Four Approaches to Measuring Organizational Strategy

Theoretical Advantages and Investigator

Issues Disadvantages Inference

Advantage Because the investigator has a

somewhat detached view, this may
be a sound method for distinguishing
between strategic changes and ad-
justments.

Strategic Change
and Adjustment Disadvantage Generally, the researcher will not
have the in-depth comparative (i.e.,
longitudinal or industry-wide) view to
allow identification of strategic
changes and adjustments. Also, the
investigator may not be granted ac-
cess to planned strategic changes or
adjustments.

Advantage This method may be more useful for

identifying intended than realized
strategies.

Intended and
Realized Strategies Disadvantage Generally, the researcher will not
have the perspective for assessing
the gap between intended and real-
ized strategies.

Advantage If the researcher is familiar with a

broad array of organizations within
the industry, this method may allow
excellent subjective assessment of
relative strategies.
The Relativity
of Strategy

Disadvantage Generally, this method will allow

only small sample sizes, thus
diminishing the opportunity for
examination of relative strategies.

534

This content downloaded from

210.57.217.22 on Fri, 11 Nov 2022 01:37:31 UTC
All use subject to https://about.jstor.org/terms
 Self-Typing Exteral Objective
Assessment Indicators

This method is ideal in its currency. Outsiders may have a co

The organization's executives are view that allows them to
most up-to-date on the organization's between strategic chang
directions. justment for a given organization. tiation between strategic changes
and strategic adjustments.

Executives may have difficulty Outsiders may not be knowledgeable This method may not reflect recent o
distinguishing between strategic or current concerning recent current changes in strategy.
changes and adjustments. strategic changes and adjustments.

This is a sound method for identify- This method is relatively well suited This method is relatively well suited
ing intended strategies. for identifying realized strategies. for identifying realized strategies. It
controls for perceptual and, to a
lesser extent, interpretive bias.

This is a questionable method for This method is relatively ill suited for This method rarely relies on data that
identifying realized strategies, identifying intended strategies. will allow identification of intended
because executives appear to strategies.
generally express strategies in terms
of intentions. And, they may express
intentions beyond those which in
fact exist.

This method allows large sample This method allows large

sizes and thus a broad view of the sizes. Each expert has a br
relative context of strategies. allowing informed rating of r
strategies. is generally possible.

Each executive's assessment of In order to ensure the experts' Data sour

strategy may have only limited familiarity with an array of organiza- limited sub
reference to the array of strategies tions, only a relatively homogeneous presentin
existing within or among industries, universe can be drawn from. pie.
Also, executives tend to perceive
their own organizations as unique.

535

This content downloaded from

210.57.217.22 on Fri, 11 Nov 2022 01:37:31 UTC
All use subject to https://about.jstor.org/terms
sometimes did not have enough information about Second, cases have arisen where members of
a competitor to make a strategic classification. In the expert panel have incomplete or outdated
part, this was a result of the geographic spread of knowledge of an organization's strategic orienta-
the sample, but more important was the fact that tion. Ideally, an investigator would be in a position
some companies operated in altogether different to select both expert panels and sample organiza-
segments of the industry. These problems led tions in a way that would maximize the ability of
Meyer [Miles & Snow, 1978, Chap. 13] to select a the experts to make their evaluations. Unfortunate-
sample of hospitals that were geographically prox- ly, this is not always possible and, when attempt-
imal and generally homogeneous in services. He ed, may result in excessive restriction of the range
then compiled a panel of experts to measure of strategic behaviors in the sample organizations.
hospital strategy that included chief hospital ad- Other things being equal, samples probably
ministrators, members of health-planning agen- should be randomly selected, which usually re-
cies, officers of local medical societies, hospital sults in incomplete ratings from experts, owing to
consultants, administrators of health-oriented limited or outdated information. To some extent,
foundations, and professors of hospital ad- this problem can be alleviated by initially selecting
ministration. Despite the wide range of back- a large number of organizations, and then, as Ham-
grounds and perspectives of the panel experts, brick [1979] did, contacting for study only those
agreement concerning the hospitals' strategies about which substantial agreement on strategy
was moderately high (.61). Hambrick [1979] also has been reached. Of course, doing so reduces the
used expert panels to measure organization randomness of the sample (but enhances con-
strategies in three different industries (health care, fidence in the measures of strategies).
higher education, and insurance). In that study, as
Objective Indicators
in Meyer's, there was general agreement within
each panel about a particular organization's A fourth measurement approach involves mea-
strategy (.49 to .76). Moreover, the expert panel pro- sures of strategy that do not rely on the percep-
cedure appeared to work well in three quite dif- tions of individuals (either internal or external to
ferent industries. the organization). An example would be published
product-market data. The main advantage of this
While the external-assessment approach ap- method is that potential perceptual biases are con-
pears to lend credence to strategies that are trolled, leaving only the more objective aspects of
measured by reliance on the investigator's obser- strategy to be measured. The disadvantage is that
vations or perceptions of the focal organization's this type of information sometimes does not exist,
managers, this approach, too, seems to have at and, even if it does, it may not be in a form that is
least two major limitations. First, in every study to usable.
date, there has been only general agreement con- This approach is commonly used by strategy
cerning strategy between the expert panels on the
researchers [e.g., Hatten & Schendel, 1977; Miller &
one hand, and the researcher and/or the organiza-
Friesen, 1978; Lenz, 1978; Datta, 1978]. To date, on-
tion's managers, on the other. Although this ly one of our studies has used this measurement
problem is not surprising, it is troubling. On occa-
approach. Hambrick [1979] focused on the central
sion, we have had to choose between these two dimension underlying the Miles and Snow typol-
groups' opinions concerning a given ogy - the rate of product-market development -
organization's strategy, and, in most instances, we
and located objective published indicators of this
have relied on the beliefs of the expert panel (on for organizations in the health care,
behavior
the presumption that these experts' views education,
were and insurance industries. He used
more objective). At the root of such discrepancies
these indicators to identify the organizations with
may be the fact that outside experts tend to report
extreme strategies (Prospectors and Defenders) in
realized strategies, whereas managers and resear-
each industry. Also, he relied on expert panels
especially selected for each industry. Finally, he
chers (if their perceptions are based on information
provided by managers) tend to report intended asked the managers to self-type their organiza-
strategies. tions. The correlations between the published ob-
536

This content downloaded from

210.57.217.22 on Fri, 11 Nov 2022 01:37:31 UTC
All use subject to https://about.jstor.org/terms
jective indicators and the assessments of strategy Conclusions
by the expert panels were .56, .46, and .41 for col-
leges, hospitals, and insurance firms, respectively Three broad conclusions about the measure-
(all significant at the .01 level). Clearly, there is ment of strategy seem warranted. First, resear-
some agreement between these objective and per- chers can enhance the validity of their strategy
ceptual measures. Nonetheless, there is substan- measures if they rely on multiple sources of infor-
tial unexplained variance, suggesting that the ap- mation. Reliance on a single data source will
proaches yield somewhat different information almost certainly yield a limited view of the
regarding the organizations' strategies. For exam- organization's current basis for competing.
ple, the objective indicators obtained from sec- Second, opportunities for strengthening the con-
ondary data sources, and the views of the expert nection between concepts and measures of
panel members, were considerably less current strategy are great. As more is learned about dif-
than the perceptions of managers. However, Ham- ferent types or aspects of strategy - for example,
brick concluded that no one approach was in- intended versus realized strategies - measures
herently superior for his purposes (linking strategy that are sensitive to these various concepts can be
to environmental scanning activities). Rather, each constructed. As our research indicates, a single
approach had costs and benefits. measure may not capture all of the subtle theo-
Hambrick's research represents the only at- retical nuances of strategy as it is currently de-
tempt we know of to employ all of the measure- fined. Finally, strategy research needs to become
ment techniques in a single study (except for the more comparative. Particularly useful would be
method of investigator inference, which has ques- studies of organizational strategies across in-
tionable validity). The results were encouraging. dustries, where the strengths and limitations of
Generally speaking, the findings suggest that, to concepts and measures would be most apparent.
the greatest possible extent, all three methods Organizational strategy is a complex and impor-
should be used in combination. Although multiple tant phenomenon that merits continued research
measures may at times yield conflicting informa- emphasis. We hope that the discussion here will
tion regarding an organization's strategy, this can serve as a point of departure for raising and analyz-
usually be reconciled and will almost always be ing additional theoretical and methodological con-
preferable to exclusive reliance on a single cerns.

measurement method.

REFERENCES

Andrews, K.R.; Learned, E.; Christensen, C.R.; & Guth,

Caves, R.E.; & Porter, M.E. From entry barriers to mobility bar-
W. Business policy: Text and cases. Homewood, III.: Irwin,riers: Conjectural decisions and contrived deterrence to new
1965.
competition. Quarterly Journal of Economics, 1977, 91,
241-261.
Andrews, K.R. The concept of corporate strat-
egy. Homewood, III.: Irwin, 1971. Chandler, A.D., Jr. Strategy and structure. New York:
Doubleday, 1962.

Beer, M.; & Davis, S.M. Creating a global organization: Clark, B.R. The organizational saga in higher education. Ad-
Failures along the way. Columbia Journal of World Business, ministrative Science Quarterly, 1972, 17, 178-184.
1976, 11, 72-84.
Cohen, M.D.; March, J.G.; & Olsen, J.P. A garbage can model
Blocker, E.E.; Bender, L.W.; & Martorana, S.V. The political of organizational choice. Administrative Science Quarterly,
terrain of American postsecondary education. Fort Lauder- 1972, 17, 1-25.
dale, Fla.: Nova University Press, 1975.
Cyert, R.M.; & March, J.G. A behavioral theory of the
firm. Englewood Cliffs, N.J.: Prentice-Hall, 1963.
Bougeois, L.J., III. Economic performance and dominant
coalition agreement on means versus ends in second-order Datta, Y. A model of competitive strategy in the TV-set in-
strategy making. In J. Susbauer (Ed.), Academy of Manage- dustry. In J. Susbauer (Ed.), Academy of Management Pro-
ment Proceedings, 1978. ceedings, 1978.

537

This content downloaded from

210.57.217.22 on Fri, 11 Nov 2022 01:37:31 UTC
All use subject to https://about.jstor.org/terms
Guth, W.D.; & Taguiri, R. Personal values and corporate Miller, D.; & Friesen, P.H. Archetypes of strategy for-
strategies. Harvard Business Review, 1965, 43, 123-132. mulation. Management Science, 1978,24, 921-933.

Hambrick, D.C. Environmental scanning, organizational Mintzberg, H. Patterns in strategy formulation. Manage-
strategy, and executive roles: A study in three industries. Un- ment Science, 1978, 24, 934-948.
published doctoral dissertation, Pennsylvania State University,
1979.
Nisbett, R.E.; & Wilson, T.D. Telling more than we know: Ver-
Hart, B.H. Strategy. New York: Praeger, 1967. bal reports on mental processes. Psychological Review, 1977,
84, 231-259.
Hatten, K.J.; & Schendel, D.E. Heterogeneity within an in-
dustry: Firm conduct in the U.S. brewing industry, Quinn, J.B. Strategic goals: Process and politics. Sloan
1952-71. Journal of Industrial Economics, 1977, 26, 97-113.
Management Review, 1977, 18, 21-37.

Hofer, C.W. Toward a contingency theory of business

strategy. Academy of Management Journal, 1975,18, 784-810. Quinn, J.B. Strategic change: "Logical Incremental-
ism." Sloan Management Review, 1978, 19, 7-21.
Hofer, C.W.; & Schendel, D. Strategy formulation: Analytical
concepts. St. Paul: West Publishing, 1978. Richards, M.D. An exploratory study of strategic failure. In
T.B. Green & D.F. Ray (Eds.), Academy of Management Pro-
Lenz, R.T. Strategic interdependence and organizational per- ceedings, 1973.
formance: Patterns in one industry. Unpublished doctoral
dissertation, Indiana University, 1978.
Snow, C.C.; & Hrebiniak, L.G. Strategy, distinctive com-
Miles, R.E.; & Snow, C.C. Organizational strategy, structure, petence, and organizational performance. Administrative
and process. New York: McGraw-Hill, 1978. Science Quarterly, 1980, 25, 317-335.

Charles C. Snow is an Associate Professor of Organiza-

tional Behavior in the College of Business Administra-
tion, Pennsylvania State University.

Donald C. Hambrick is an Assistant Professor of Busi-

ness in the Graduate School of Business, Columbia
University.

Received 7/9/79

538

This content downloaded from

210.57.217.22 on Fri, 11 Nov 2022 01:37:31 UTC
All use subject to https://about.jstor.org/terms